This required consistency can be achieved by the normal "shock" graining technique only when operated by expert pan boilers in conjunction with a regular steam supply, independent vacuum control and some type of pan control instumrent.
However, when a " t r u e " seeding technique is used in conjunction with a pan control instrument, the first three requirements are somewhat modified. The pan boiler's judgement on quantity of grain is entirely eliminated, the steam supply may fluctuate within reasonable limits and, in the case of conduc- tivity pan contol at least, the effect of vacuum fluctuations can be minimized to a large extent by holding a steady conductivity value. However, to achieve this control over the effect of vacuum fluc- tuations, the pan boiler must be extremely watchful, especially during the initial growth of the crystals, when they are very far apart and the risk of secon- dary grain formation is very real.
A corollary to the elimination of the pan boiler's judgement in graining, is that the responsibility for the final crystal size moves from the pan boiler to the Factory Manager.
The " T r u e " Seeding Technique
The requirment for " t r u e " seeding, is t h a t the number of sugar particles drawn into the pan at graining should equal the number of sugar crystals in the finished stike. (It follows that by controlling the number of crystals present, the final crystal size is also fixed for any given set of conditions)
There are two methods of achieving this:—the dry powder method developed by Alewijn and Honig1 in Java and used in Natal by Steyn2, and the slurry method, developed along slightly different lines by Gillett3 in America, and Pollard4 and Cameron5 in Australia.
The former method is fully described in S.M.R.I.
Tech. Comm. No. 9 and by Steyn in a paper at the present Conference. Steyn's figures agree closely with those of Honig and Alewijn.
The American slurry method, consits of mixing a very fine commercial powdered sugar into a slurry with some sugar insoluble liquid, such as isopropyl alcohol. Unfortunately this powdred sugar is not readily obtainable outside the U.S.A. and so severely limits the use of this method.
The Australian practice on the other hand, overcomes this difficulty by grinding refined sugar in methylated spirits, in a high speed vertical ball mill.
In either method, whether dry powder or slurry is used for seeding, the basic boiling pattern is the same.
The graining charge is boiled down until a small degree of oversaturation is obtained, at which point
the seed is introduced. Balancing water is then used to hold the charge close to this supersaturation value until the grain has grown to sufficient, size to safely take syrup or molasses feed. Hereafter the super- saturation is increased to a normal boiling value with the required feed (syrup or molasses).
During the past 1953 54 crushing season, the S.M.R.I. carried out a series of preliminary tests with slurry graining at Umfolozi.
The following is a. brief report on this work and some tentative conclusions are reached.
1. Slurry Preparation.
The slurry was prepared in a horizontal pebble mill of internal dimensions, 5½" long by 7½" diameter.
The apparent, volume of the pebbles half filled the.
mill, which was driven at 60 r.p.m.
1220 gms, of Hulett's refined sugar and 1500 ml 95 per cent, alcohol were ground in this mill for 7½ hours, and then made up to 1 gallon and divided equally into six bottles. One slurry batch was made with methylated spirits instead of 95 per cent, alcohol.
2. The Graining Tests.
As these tests could only be regarded as pre- liminary trials into the applicability of the method to South African conditions, and the Umfolozi pan stage is usually severely pressed for capacity, it. was decided to follow the normal Umfolozi boiling procedure, modifying it only to take slurry seeding.
This normal boiling procedure is:— to shock- grain a blend of 80° purity in No. 5 pan, build the pan up to 1,100 cu. ft. and 74° purity, and cut two thirds to a seed vessel. This makes one C strike in No. 5 pan and two in No. 7, giving a total of about 2,300 cu. ft. of C massecuite.
The modification merely consisted of boiling down the graining charge to just over saturation, noting the conductivity reading, and holding it at this reading with hot balancing water. When this was achieved, three bottles of slurry were sucked into the pan, via a funnel and graining cock, and the pan held on balancing water until the grain had grown sufficiently for syrup feed to be started. The opera- tion usually took about 40 minutes from graining to start of syrup feed. Boiling was then continued in
the normal manner.
Unfortunately calandria steam pressure cannot be given for these strikes due to No. 5 pan lacking a pressure guage. However, the plan followed was to vigorously boil the charge down to the graining point, then cut the steam back to give gentle boiling and gradually increase it while using water feed.
Full boiling rate being used again at the s t a r t of syrup feed.
All the strikes were boiled with conductivity control. The aim being to keep these values as steady as possible notwithstanding the effect of vacuum fluctuations. These fluctuations can be quite severe at Umfolozi, where six pans all work off the same vacuum line.
Typical conductivity readings for these seed strikes were:
(1) Graining point = 51 mA.
(2) Syrup feed started = 48 mA.
(3) Pan full = 43 mA.
In all, six grainings were carried out: - 4 in No. 5 pan using ethanol slurry.
1 in No. 5 pan using methylated slurry.
1 in No. 7 pan using methylated slurry.
The first attempt in No. 5 resulted in a very poor grain. This was due to the pan being heavied up too much immediately after syrup feed was started, and resulted in a second crop. This was washed out, but the washing caused a high percentage of conglome- rate of form. However, this had the value of defining the conductivity limits. No measurements of the resulting grain could be obtained.
The other three grainings with ethanol slurry, were most satisfactory from the consistency point of view. The finished seed grain sizes were 0.222, 0.221 and 0.233 m.m and the percentage deformed crystals was 20.5, 21.0 and 25.6 per cent, respectively.
These latter figures m a y appear rather high and will be discussed later in this paper.
The fifth graining in No. 5, and t h a t in No. 7, were both carried out with the same batch of methylated slurry, and therefore, even though only one graining is being examined in each case, a certain amount of significance can be attached to the results.
The finished seed in No. 5 showed an extremely high percentage of deformed crystals, viz. 54.7 per cent., and so rendered the size determination most unsatisfactory. However, this appeared to be about 0.273 mm., i.e. considerably higher t h a n with the ethanol slurry grainings. Against, these figures, the graining in No. 7 pan gave only 10 per cent, deformi- ties and the large crystal size of 0.362 mm.
3. Conclusions from the Umfolozi Tests No definite conclusions can be drawn from the results at Umfolozi, due to the small number of results obtained, however t h e y do indicate several important points.
1. Conductivity readings may be used to re- produce the graining point, even at high purities—
all graingings in No. 5 pan were effected between 50 and 54 mA., even though one series was carried out in October and the other in January.
2. Crystal size reproducibility between strikes is good, providing the slurries are made up in the identical manner with the same materials.
3. No. 5 pan has poor boiling properties when compared with No. 7, as is shown by the deformed crystal percentages recorded for both pans.
This was expected from the design of the two pans.
No. 5 pan is of very poor design, with a high floating calandria and fairly steep conical bottom, giving a large dead volume under the calandria. In the original design, coils were installed in this dead volume to assist the circulation, however, at some time during their life, these coils became leaky and were removed, a procedure which must have materi- ally affected the performance of the pan. Added to this, a belt in the pan wall was installed to increase the stike volume from 750 ± cu. ft. to 1100 ± cu. ft.
This increase in strike volume is suspected to account, to some extent, for the high deformity percentage recorded in the methylated slurry graining in this pan.
In this case the strike was carried up slightly lighter than the others, and, while the proofs taken during the stike appeared to be quite normal, the finished strike showed a very high percentage of deformities.
The question arising from this is:—-was the super- saturation at the end of the stike sufficiently low for the effect of hydrostatic head to cause undersatu- ration at the calandria, with resultant crystal dissolution and hence conglomerate formation?
Unfortunately this was the second to last strike boiled in the series, and, while photographs were taken to the grain at various stages of the next strike to enable deformity counts to be taken, the strike was run up somewhat heavier than the other and showed a reasonably constant value of about 20 per cent, deformities throughout its life.
No. 7 pan on the other hand, is of modern "centre flow" design with, excellent boiling properties, however, the graining volume is very high, being 40 per cent, of the normal pan capacity. This means t h a t the grain is very widely separated at the beginning of the strike and so the pan must be handled most delicately at this stage.
4. Slurry graining in conjunction with a con- ductivity instrument largely minimizes the effect of vacuum fluctuations on graining.
Discussion
Two methods of "true." seeding have been, men- tioned in the paper, one with dust and the other with slurry. Unfortunately it is impossible to say at present which is the better method, for to determine this, one must take into account, the distribution of crystal, size about the mean, as well as the repro- ducibility between strikes, and too few analyses showing this are to hand as yet.
Further work on this subject is planned for the coming crushing season.
Acknowledgement
This Institute wishes to thank the Management and Staff of Umlolozi Sugar Mill for the assistance and facilities provided for this work.
R E F E R E N C E S
1 Alewjin, W. F., Archief v.d. J.S.I., No. 11, 1933, 334.
2 Steyn, C. L., Proceedings S.A.S.T.A., 1954.
3 Gillett, E. C., Low Grade Sugar Crystallisation, 1948.
4 Pollard, J. S., "An Improved Method of Graining," Proceed- ings Queensland S.S.C.T., 1951, 55.
5 Cameron, E. B. G., Private Communication, 1952.
(For discussion on this paper see page 119.